A complete application in a minute
Tailored resume and cover letter, ready to send.
Training Turk is seeking an experienced electrical engineer to design and evaluate circuits and boards for training large language models in electrical‑engineering tasks. You will create domain‑specific problems and author precise reference solutions covering circuits, power, and embedded hardware.
Responsibilities include scoring AI‑generated results, building robust rubrics for reasoning across EE sub‑domains, and collaborating with SMEs to ensure data quality.
Design and evaluate electrical‑engineering problems to train large language models.
Circuit design Power systems Embedded hardware
“Describe how you would construct a prompt for an LLM to design a buck‑converter that must meet 95% efficiency at 12V input, given a 3‑week deadline and a target cost under $5 per unit.”
Task design for AI training
“Describe how you would construct a prompt for an LLM to design a buck‑converter that must meet 95% efficiency at 12V input, given a 3‑week deadline and a target cost under $5 per unit.”
Why this comes up
They need to confirm you can translate real engineering challenges into clear, solvable prompts that stress model capabilities.
“Explain the detailed calculation steps you would include in a reference solution for sizing a heat sink for a 150W power amplifier operating at 85°C ambient.”
Reference solution authoring
“Explain the detailed calculation steps you would include in a reference solution for sizing a heat sink for a 150W power amplifier operating at 85°C ambient.”
Why this comes up
Your ability to produce precise, step‑by‑step solutions demonstrates the benchmark quality the model will learn from.
“A model suggests using a 10µF capacitor in a high‑frequency RF filter; outline the red flags you would raise and the correct approach.”
Model output critique
“A model suggests using a 10µF capacitor in a high‑frequency RF filter; outline the red flags you would raise and the correct approach.”
Why this comes up
Interviewers assess whether you can quickly identify subtle engineering mistakes that an AI might make.
Rubric development
“Draft the key rubric items you would use to grade an AI‑generated PCB layout for a 4‑layer board handling 5A current.”
Rubric development
“Draft the key rubric items you would use to grade an AI‑generated PCB layout for a 4‑layer board handling 5A current.”
Why this comes up
Creating consistent evaluation criteria ensures reliable scoring across many reviewers and tasks.
Collaboration with SMEs
“You notice a discrepancy between your power‑distribution task and a colleague’s control‑systems task; how would you resolve it before the data is finalized?”
Collaboration with SMEs
“You notice a discrepancy between your power‑distribution task and a colleague’s control‑systems task; how would you resolve it before the data is finalized?”
Why this comes up
Coordinating with other experts tests your ability to maintain data quality and avoid contradictory guidance.
Candidates may be asked to design a realistic power‑electronics problem, write a full solution, and create a rubric to evaluate AI‑generated answers within a limited time frame.
Pay $80-120/hr
Hours Full time, 40 hours a week
Where Remote, United States
Open to USA
Field Other Engineering
Posted 10/15/2026
Places left 20
We wrote this page from the public Mercor listing. It may be out of date, so read the full posting before you apply.